I have been troubleshooting months of system-wide stuttering on a Ryzen 9 9950X3D, ASUS ROG STRIX X870E-H GAMING WIFI7, 64GB DDR5-6000, and RTX 5060 Ti system. The issue persists across several BIOS versions and even after a completely clean Windows 11 Pro 25H2 installation. It affects window dragging, DWM animations, typing in multiple browsers, YouTube playback startup, and game frame pacing despite high average FPS. Typing can pause every 8–15 characters while one or two CPU threads briefly spike, and the fans sometimes ramp up during those pauses. Windows also briefly loses display signal during boot before the login screen appears.
LatencyMon previously showed nvlddmkm.sys at roughly 3500 µs, while the clean installation usually points to ACPI.sys at around 1500 µs. OCCT, TestMem5, and CPU tests show no errors or WHEA events. The only instability found was an overly aggressive all-core Curve Optimizer setting, so the current per-CCD settings are -15 and -5.
I have tested multiple NVIDIA and chipset drivers, DDU-based driver installations, BIOS defaults, EXPO and manual memory settings, different power plans, HAGS, MPO, Game Mode, browser hardware acceleration, USB ports, mouse polling rates, display cables, Adaptive Sync, G-Sync, V-Sync, PCIe and ASPM settings, C-states, PBO, and various motherboard power options. The GPU worked normally in my previous Intel system and has shown no artifacts, crashes, driver resets, or black screens. Interestingly, NVIDIA driver 596.21 feels smoother than newer versions, and mostly automatic BIOS settings seem to behave best.
Could this be an AMD platform, motherboard firmware, ACPI, PCIe, USB, chipset, scheduling, or Windows latency problem rather than a faulty GPU? I am also considering testing a Linux live USB to determine whether the behavior exists outside Windows.
2 Answers
Since the graphics card behaved normally in the previous system, it is less likely to be the primary fault. The new processor also produces considerably more heat than the old i7-8700, so cooling and CPU temperature behavior are worth checking carefully, even though the current liquid cooler should normally be capable of handling it.
A problem that survives different drivers, BIOS versions, and a clean Windows installation could still be related to UEFI or the platform itself. A useful isolation test would be a Linux live USB: check whether window movement, typing, video playback, and general responsiveness stutter there too. Also test with the firmware completely at defaults, including no EXPO, PBO, or Curve Optimizer, because memory and boost settings can create latency without producing crashes.
The stutter also occurs with the BIOS at default settings. I am planning to create a live USB and compare the behavior outside Windows.
The old system used a Thermalright Peerless Assassin 120, while the new system uses an Arctic Liquid Freezer III 360 Pro. Cooling was adequate on the previous system, and I have not seen obvious temperature-related instability.